Technology is changing the way sports content is consumed, but it is data that is revolutionising the way in which that content is understood. If sports fandom is all about debate and perspective, forming opinions and making sense of the action with the help of data is, at least for the geeks who capture the stuff, about contextualisation and visualisation.
For Sportradar, the leading sports data and betting services company, the need to crunch numbers into narrative has never been greater. The sheer amount of raw data now available has meant the company has never had such a wealth of information at its disposal - today, Sportradar supplies sports data to over 1,000 media companies, bookmakers, sports federations and government authorities in over 80 countries. Yet the ability to distil that data down into engaging, digestible content is where the power of the sports data company really shines through.
“Raw data is so widely available it’s pretty much become a commodity these days,” says Ashok Balakrishnan, Sportradar’s senior vice president for product and technology. “Just having access to the data is no longer a competitive advantage for us as a company. It’s how do you contextualise it, add a more semantic layer that truly makes it a lot richer? How do you turn data into information into knowledge?”
It is that quest to engage fans and tell more compelling stories through data that has inspired Sportradar’s latest acquisition. On Thursday, the US division of the Swiss-headquartered company announced its purchase of MOCAP Analytics, a Silicon Valley startup that specialises in artificial intelligence (AI)-powered player-tracking technology.
Co-founded just five years ago by Arian Forouhar and Eldar Akhmetgaliyev, MOCAP Analytics is perhaps best known in sport for its work with the Golden State Warriors, the all-conquering National Basketball Association (NBA) champions who became the first team to use the company’s proprietary technology in 2012. The firm is based in Menlo Park, California, right at the heart of America’s booming tech industry, but it was less its location than its software and expertise that caught the attention of Sportradar.
Having spent the past few years branching out from its origins in the provision of data and betting services, Sportradar has designs on becoming a full-service content company offering an array of digital data products specifically designed to drive fan engagement. To that end, it has struck up official relationships with a host of leading sports and media entities - including in the United States with the National Football League (NFL) and the NBA - and has also recently expanded into the over-the-top (OTT) streaming space following its acquisition last year of the Sportsman Media Group.
MOCAP's AI technology will be applied to Game Stream, Sportradar's new NFL player-tracking and play visualisation software.
“When coming up with our vision and strategy for the US market, it was quite evident that there was a need for AI capabilities,” Balakrishnan, who joined Sportradar in March and is now based out of the company’s New York office, tells SportsPro. “For us a company, it’s not an option - it’s definitely an imperative.
“We literally scoured the entire industry and we found MOCAP to be the only company that had experience of six years. There are obviously a lot of other companies that are all trying to dabble in the space but they’re all very recent. When we looked at MOCAP through our research and met them and saw under the covers, we got a full insight into what they had built. For us it was apparent that it was the right choice.”
For MOCAP - a company that, to quote Balakrishnan, comprises “a bunch of PhDs and hardcore geeks into their technology” - the benefits of coming under the banner of Sportradar are clear. “From their perspective, they see a massive opportunity to unlock the potential of the technology that they’ve built and scale it,” says Balakrishnan, whose company has grown to employ over 1,700 people across offices in 34 cities worldwide. “They’re probably not the best from a business development and sales perspective. That’s not their strength so that’s where they see huge potential of partnering with someone like us with our network, being able to bring their technology to widespread customers.”
Following their integration into the Sportradar business, MOCAP's workforce will eventually form a newly created Analytics & Storytelling division headed by Forouhar, the company’s chief executive who, like Akhmetgaliyev, holds a PhD from the California Institute of Technology. According to Balakrishnan, the unit’s overarching remit will be to develop new fan-centric data products that will then be made available through Sportradar's industry clients including broadcasters, digital publishers, online distribution outlets, brands, teams and leagues.
What we’re trying to do is take that mountain of data and then apply more advanced techniques using AI and machine learning to come up with insights and storylines that are really engaging for fans.
“With those key verticals in mind,” he adds, “what we’re also doing is we’re not going off and building one-off products for each one of those. We’re starting on building what I would call an advanced technology platform with key capabilities we can use as building blocks to assemble in many different ways, to come up with products for each of those key verticals.
“If you look at AI being a key capability, we’re also building another capability called Natural Language Generation (NLG), which allows all the data insights to be used in the form of automated textual generation that come up with these storylines as a feed or something that goes into a broadcaster tool on TV.
“Also, fully-automated content generation pipelines, where you have all of the insights coming out of our AI layer going through NLG, then feeding into an automated content generation so a 30 or 60-second video clip can be generated with all these rich, insightful storylines, and that can go out to social media.”
Balakrishnan explains that MOCAP’s machine learning software and algorithms will initially be applied to Game Stream, a new football-focused product rolled out this September and which provides bite-sized, interactive visualisations that leverage the NFL's Next Gen Stats player-tracking data - gathered from sensors in players’ helmets and distributed by Sportradar.
The Game Stream product, which is intended for use on so-called ‘second screens’, allows fans to access statistical information pertaining to key plays that was never available before in real-time, such as the speed at which a player is running, the time taken for a quarterback to throw a pass, or the distance between the wide receiver and the tackler at the moment it was thrown or caught. Now, MOCAP’s AI-generated content will be laid over that product, offering play projections and providing further context to the data on screen.
Using MOCAP's software, Game Stream will now provide play projections in the form of textual analysis.
Other Sportradar clients will benefit in similar ways, says Balakrishnan. Broadcasters, for example, will have the ability to use MOCAP’s technology to create on-screen graphics for use before, during and after games, while on-air analysts will be able to use the company’s content to inform their previews and reviews.
For the NBA, too, Balakrishnan says harnessing MOCAP's technology will enhance Sportradar’s existing offerings. As the official data partner and real-time statistics provider of the league, the company has exclusive access to a wealth of player-tracking data provided by LA-based Second Spectrum, effectively a MOCAP rival that has installed its own camera-based tracking systems in every NBA arena as part of a wide-ranging partnership agreed in September 2016.
Sportradar will continue to have access to that data, but now it will be able to run it through MOCAP’s algorithms, offering media clients an enhanced set of products and solutions. “It’s not just about giving the statistics and box score data to the rest of the world,” says Balakrishnan. “It’s about, how do we truly help come up with a new storytelling experience around the game and the player performance and all the previews of the match-up?
“That’s where the AI layer that MOCAP has built will help us come up with premium products around that space. We will then go and sell that separately as a much more premium product to all of our customers. They might get the raw data feed from us which is baseline and basic, and then through MOCAP they can get a more enriched set of data.”
Ultimately, adds Balakrishnan, the aim is to take data-driven sports content to the next level by developing advanced visualisation solutions. He argues that while the amount of data available has exploded in general in recent times, the quality of that data has in fact “plateaued” over the past ten or 15 years, with many stakeholders in sport rarely looking beyond the basic, albeit abundant, box score and play-by-play statistics.
“What we’re trying to do is take that mountain of data and then apply more advanced techniques using AI and machine learning to come up with insights and storylines that are really engaging for fans,” he explains.
“When we do that we want to cater to the fan throughout the different aspects of their fan journey. Whether they’re watching the game on the television or on a digital platform or OTT environment or at a venue or accessing content on a digital publisher, reading articles, we want to be there serving them with the most engaging storytelling content.”